Background: In 2008, the American Diabetes Association (ADA) suggested metformin be used in patients with diabetes and reduced renal function. We evaluated time trends in metformin discontinuation for patients who reached reduced renal threshold.

Methods: Using National Veterans Administration and Medicare databases, we assembled a cohort of metformin and glipizide users who reached serum creatinine >1.5/1.4 mg/dL for males/females. Odds ratio (OR) of medication discontinuation within 180 days of renal threshold over time was modeled using multivariable logistic regression, with metformin discontinuation in 2008 as reference.

Results: There were 18,025 metformin patients and 22,595 glipizide who met renal threshold. In 2008, 55.3% metformin patients discontinued within 180 days of renal threshold. Figure 1 depicts probability of metformin and glipizide discontinuation annually. The odds of metformin discontinuation was 1.25 (1.23, 1.35) vs. glipizide. In a fitted regression of discontinuation probability, there was no statistical difference in slopes between the two treatments (p=0.125).

Conclusions: Probability of discontinuing metformin with reduced renal function decreased over time but was more likely than glipizide discontinuation. The decrease in discontinuation over time was similar in both treatments.


P. Chu: None. M.R. Griffin: None. J. Min: None. A.J. Hackstadt: None. J.J. Chipman: None. R.A. Greevy: None. A.M. Hung: None. C.G. Grijalva: None. C. Roumie: None.


U.S. Department of Veterans Affairs (CX000570-07 to C.R.)

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at